Digital Signal Processing for The Development of Deep Learning-Based Speech Recognition Technology

Abstract

This research discusses digital signal processing in the context of developing deep learning-based speech recognition technology. Given the increasing demand for accurate and efficient speech recognition systems, digital signal processing techniques are essential. The research method used is an experimental method with a quantitative approach. This research method consists of several stages: introduction, research design, data collection, data preprocessing, Deep Learning Model Development, performance training and evaluation, experiments and testing, and data analysis. These findings are expected to contribute to developing more sophisticated and applicable speech recognition systems in various fields. For example, in virtual assistants such as Siri and Google Assistant, improved speech recognition accuracy will allow for more natural interactions and faster responses, improving the user experience. This technology can be used in security systems for safer and more reliable voice authentication, replacing or supplementing passwords and fingerprints. Additionally, in accessibility technology, more accurate voice recognition will be particularly beneficial for individuals with visual impairments or mobility, allowing them to control devices and access information with just voice commands. Other benefits include improvements in automated phone apps, automatic transcription for meetings or conferences, and the development of smart home devices that can be fully voice-operated.

Keywords

• digital signal processing
• deep learning
• speech recognition
• technology.